Ellen Tsaprailis, April 22, 2022
Photo credit: Lindsay Ralph
Ericsson Fellow Uses Machine Learning to Reduce Delays in 5G Real-Life Scenarios
Taking a break from working in industry to join Carleton as an Ericsson Fellow and pursue a master’s degree was the right move for Vishnu Priya Guddeti to get the opportunity to specialize in a research area she is passionate about.
Guddeti is researching latency reduction for IoT applications in 5G networks. Her research investigates Machine Learning approaches to optimize processing of data generated from static and mobile IoT devices in an edge cloud environment. Looking at ways to reduce delays in different 5G applications, examples include automated traffic signals that sense when a pedestrian is ready to cross the street or minimizing the time it takes for face recognition to open cellphones in microseconds.
“A face recognition system, for example, should be able to recognize our faces quickly,” says Guddeti. “Whenever we open the app, it should recognize and give access without delays. I work on how you reduce computation delays with 5G networks using cloud-edge-computing and come up with novel concepts that work for different real-life scenarios.”
Guddeti is one of six graduate students who are Ericsson Fellows at Carleton University—a unique, talent-building program born out of the Ericsson-Carleton University Partnership for Research and Leadership in Wireless Networks.
Instead of working as a teaching assistant during their graduate studies, Guddeti and the other fellows are being supported to focus on their pioneering wireless communications research and get input from both their academic supervisors and Ericsson professionals.
Having graduated in 2017 with an undergraduate degree in electrical and electronics engineering from Vellore Institute of Technology in India, Guddeti started working as a junior engineer at Ford Motor Company on advanced driver assistance systems.
Having developed a keen interest in Intelligent Navigation Systems, Guddeti looked for a way to specialize so she decided to apply for a master’s in Applied Science where she could work on concepts in-depth while completing her studies.
Research projects carried out in the Embedded and Multi-sensor Systems Lab (EMSLab) at Carleton University were closely related to the work she was interested in and the Master of Applied Science in Systems and Computer Engineering seemed like a tailor-made program. Guddeti was pleased to be admitted into the MASc and offered a research assistant position in the lab of Associate Professor and EMSLab Director Mohamed Atia.
“The Fourth Industrial Revolution that is currently evolving and shaping our 21st century is marked by two key aspects: connectivity and autonomy,” says Professor Atia.
“Connectivity is currently evolving from the fourth-generation networks that enabled broadband mobile communication to the fifth-generation networks (5G) that enables high bandwidth: 100 times the number of connected devices and 100 per cent coverage. This better connectivity solves computation problems by enabling distributed processing of data to applications such as traffic monitoring, the Internet of Things, augmented reality and autonomous vehicles. I was fortunate to be the academic supervisor of Vishnu who is supported by the Ericsson Fellowship. Vishnu’s work is transformative research that applies advanced Machine Learning algorithms to optimally distribute computation over heterogeneous resources including local devices, edge servers and cloud servers.”
Guddeti will defend her thesis this month and her career plan is to gain more professional experience in the field of smart mobility and connected cars.
“The research I have done until now and the concepts I’ve learned have given me a good background to further develop these skills,” says Guddeti.
Research Papers Submitted for Peer Review
Guddeti has written two papers already submitted for peer review and is on her way to creating impact in the 5G network.
One paper is titled, Offloading in Edge Cloud Environment using Deep Q Reinforcement Learning which has been submitted to the journal IEEE Transactions on Emerging Topics in Computational Intelligence.
“This paper is about reducing delay, and utilized bandwidth to support emerging IoT applications in edge computing,” says Guddeti. “A novel deep reinforcement learning (Machine Learning) algorithm is proposed and validated through extensive simulations for real-time data.”
The second paper is titled, Decentralized Offloading and Container-Based Caching in Edge Computing which has been submitted to the journal IEEE Internet of Things.
“This paper is an extension of the first paper and includes the advantages of caching application-related data and analyzes the deployment location of algorithm,” says Guddeti.
In this prestigious fellowship program, Carleton graduate students conduct hands-on research alongside Ericsson experts in state-of-the-art facilities, ensuring students build skills that are in high demand in today’s telecommunications industry.
Share: Twitter, Facebook